As the US midterm elections bring changes to the regulatory environment surrounding data center development, attention is turning to whether demand will expand for technologies that improve the power efficiency of AI data centers.
According to Reuters, the US Democratic Party secured majorities in both the Senate and the House of Representatives in the 2026 midterm elections. At the same time, local opposition to data centers' electricity and water use and environmental impacts is growing across the United States. Most data center-related local ballot measures held alongside the elections called for development restrictions or additional regulation.
These developments could affect permitting, construction schedules and investment incentives for new data centers. In major data center investment regions, including Texas, there are moves to temporarily suspend new permits or tighten regulation of electricity costs and environmental impacts.
However, even if new data center construction slows, electricity demand associated with expanding AI computing is likely to continue. Power supply pressures are increasing in the US PJM grid, where electricity demand is expected to grow by up to 30 GW by 2030, partly because of the increase in AI data centers.
This is increasing the importance of high-efficiency power infrastructure technologies that reduce power losses at existing facilities and allow more computing equipment to operate within the same power capacity.
Against this backdrop, SOLUM is developing high-voltage power solutions targeting AI data centers. The company has developed a 26 kW-class DC-DC converter for battery backup units (BBUs), based on 800 Vdc high-voltage direct current, and is pursuing entry into the global hyperscale data center supply chain.
SOLUM completed sample validation of the product in the first quarter of 2026 and aims to begin mass production in the first quarter of 2027 following further compatibility testing and customer validation.
The 800 Vdc approach can transmit the same amount of power at a lower current, making it advantageous for reducing power losses and heat generation during distribution. As increasingly dense AI servers raise power consumption per rack, power conversion efficiency and improvements to distribution architecture are considered factors that determine data center operating efficiency.
Building on power conversion technology accumulated through its existing cloud server power supply business, SOLUM is expanding into power products for AI data centers, including BBUs and high-voltage DC-DC converters. The company is currently conducting further validation procedures to enter global hyperscalers' supply chains.
However, if restrictions on new data center construction expand in the United States, new orders for server racks and power equipment could be delayed. Conversely, continued upgrades to existing data center power infrastructure and the transition of next-generation AI servers to high-voltage power architectures could generate new demand for companies with relevant power conversion technologies.
The results of SOLUM's AI data center business are expected to become clearer depending on future customer technology validation, mass production approval and whether actual supply contracts are signed.
Original Korean article: dt.co.kr
English brief by Korea Wind Intelligence, koreawindintel.com
